Last week, I did something I've been wanting to do for months: I stayed in Claude while working with my company's data.
Here's what happened:
- I had a PDF with updated business plan metrics
- I needed to add it to our Kairos docs
- Then generate a Canva presentation from those updated numbers
- All without leaving Claude
No copy-paste. No switching between tabs. No manual data entry.
I just told Claude what I needed, and it read from my Kairos workspace, modified the data, and created the presentation—all through our new MCP connector.
This is the workflow consolidation problem we've been talking about. But solved at a different layer.
What's an MCP? (And Why Should You Care?)
MCP stands for Model Context Protocol. If you haven't heard of it, here's the simplest explanation:
It's a way to give Claude (or any AI) direct access to your apps and data.
Without MCP:
- You use Claude in isolation
- You copy-paste data in
- Claude generates output
- You copy-paste it into another app
With MCP (like our Kairos connector):
- Claude connects directly to your Kairos workspace
- Claude reads your actual tasks, docs, goals
- Claude understands your real context
- Claude writes directly back to Kairos
- You stay in Claude the entire time
The difference: AI that knows your work vs. AI that guesses.
Why This Matters (More Than It Seems)
Most AI tools are isolated. They're smart, but they don't know anything about your actual work.
You ask Claude to "prioritize my tasks," and Claude has to guess. Are they product tasks? Support tasks? Roadmap items? What's the context?
With the Kairos MCP, Claude knows your tasks. Claude sees your goals, your roadmap, your team's actual work. Claude's suggestions aren't generic—they're grounded in reality.
This is the difference between:
- Generic AI: "Here are 5 ways to prioritize tasks"
- Contextual AI: "Based on your Q1 goals and current sprint, here's what should move up"
One is helpful. The other changes how you work.
Our Real Use Case: PDF → Kairos → Canva
Here's exactly what I did (and how you can too):
The Setup
I had:
- A PDF with updated business metrics (in Google Drive)
- An outdated Kairos doc with our company overview
- A need to create a Canva presentation for investors
- Zero desire to manually copy data between three apps
The Workflow (What Actually Happened)
Step 1: Upload PDF to Kairos via Claude
Step 2: Generate Canva Presentation from Kairos Data
What I saved:
- 30 minutes of manual copy-paste
- 3 context switches (Google Drive → Kairos → Canva)
- The risk of entering data incorrectly
- The cognitive load of "where did this number go?"
What Claude saved:
- Having to ask me 10 clarifying questions about outdated data
- Working with stale information
- Generic suggestions vs. actually useful ones
How the Kairos MCP Works
You can access Kairos from three places:
1. Claude.ai (Web)
- Log in to Claude.ai
- Enable the Kairos MCP connector in settings
- Start using Kairos directly in chat
- "Read my Q1 roadmap," "Create a task," "What's blocking us?"
2. Claude Code (Terminal/Local)
- Install Claude Code
- Configure the Kairos MCP
- Use Kairos data in your scripts/automation
- "Generate a report from my tasks," "Auto-create issues based on feedback"
3. Claude Desktop (Mac/Windows)
- Install Claude Desktop
- Add Kairos MCP to configuration
- Same access, native app experience
- "Summarize my goals," "What's my team working on?"
All three have the same capabilities: read tasks, docs, goals, chat; write new items; update existing data.
What You Can Actually Do (Real Examples)
Once you connect Kairos MCP, the possibilities expand fast.
Example 1: Meeting Prep
Claude reads Kairos, gives you talking points instantly.
Example 2: Async Status Updates
Claude pulls completed tasks from Kairos, formats it for sharing.
Example 3: Customer Feedback → Tasks
Claude parses the email, creates structured tasks, links them to goals.
Example 4: Roadmap Analysis
Claude reads your roadmap, spots conflicts, suggests consolidations.
Example 5: Documentation Generation
Claude creates structured docs directly in your workspace.
Example 6: Report Generation (Like I Did)
Claude reads Kairos, generates a presentation, shares the link.
Why We Built This (And Why It Matters)
We built the Kairos MCP for one reason: to eliminate the gap between thinking and doing.
Most teams have this workflow:
- Think about work (in Claude, Slack, a meeting)
- Manually create it in your tool (Kairos, Jira, whatever)
- Reference it later (search through 5 different places)
The MCP removes step 2. Claude is your tool.
You think → Claude creates → It's already in your workspace.
This is what "AI-native workspace" actually means. Not "AI sprinkled on top of your tools." But AI that is your operating system.
How to Get Started
Step 1: Update to Latest Kairos
Make sure you're on the latest version. MCP requires recent infrastructure.
Step 2: Enable in Claude
- Go to Claude.ai or Claude Desktop settings
- Find "Connected apps" or "MCP servers"
- Search for "Kairos Connect"
- Click "Enable"
- Authorize your Kairos account
Step 3: Start Using It
Just ask Claude what you need:
- "What's in my Kairos?"
- "Create a task for..."
- "Update my Q1 goals"
- "Generate a report from..."
That's it. Claude handles the rest.
Step 4: Build Workflows
Once you understand the basics, build workflows:
- Read from Kairos → Generate content in Canva
- Read customer feedback → Create Kairos tasks
- Read your roadmap → Generate presentations
- Anything you can describe, Claude can automate
Technical Details (For Developers)
If you're building on top of Kairos:
MCP Capabilities:
- Read tasks, docs, goals, chat messages
- Create new tasks, docs, goals
- Update existing items
- Query across your workspace
- Real-time sync with Kairos data
Available in:
- Claude.ai (web)
- Claude Code (CLI)
- Claude Desktop (native apps)
- Any tool that supports MCP protocol
Authentication:
- OAuth via your Kairos account
- Permissions based on your Kairos workspace access
- No API keys needed (safer for Claude)
Rate Limits:
- Standard Claude rate limits apply
- Kairos API calls included in your plan
- No additional charges for MCP usage
Documentation:
- Full API reference: docs.kairos.app/mcp
- Code examples: github.com/kairos/mcp-examples
- Community discussion: Discord
Why This Changes Everything (For Our ICP)
If you're a founder or CTO managing a team:
Before MCP:
- You use Claude for thinking/writing
- You manually update Kairos with decisions
- Your team checks Kairos separately for context
- AI suggestions are generic
After MCP:
- You stay in Claude (your thinking tool)
- Claude auto-updates Kairos (your work tool)
- Your team sees real-time updates automatically
- AI suggestions are grounded in your actual work
The math:
- 1-2 hours/week saved on "syncing" work between systems
- Better AI suggestions (contextual vs. generic)
- Less context switching
- Your workspace becomes truly AI-native
The Future (Why We're Excited About This)
This is just the beginning.
In a few months, we're adding:
- Automated workflows – "Every Friday, generate a status report from completed tasks"
- Integration chains – "Update Kairos task → Auto-sync to GitHub issue → Notify Slack"
- AI-powered suggestions – Claude suggests next steps based on your work patterns
- Custom MCP extensions – Build your own Claude ↔ Kairos workflows
The vision: A workspace where AI isn't a separate tool. It's part of how you work.
Questions? Try It Out.
The Kairos MCP is live now. You can:
- Enable it today – Takes 2 minutes
- Try it with your data – See what's possible
- Share feedback – What workflows would help you?
- Join the discussion – Discord community
Or reach out if you have questions.
This is new territory. We're learning what's possible with AI + data + workspace integration. Your feedback shapes what we build next.
What workflow would you automate if Claude had access to your work data? Drop a comment or let me know—I'd love to hear what you'd build.